Computer systems not only store data to be accessed both locally and through a network but they are also capable of performing a wide variety of analytics relating to the data stored within the system. Analytics software suites can process raw network data to provide valuable information to network administrators or other interested parties related to network usage, network efficiencies, network bottlenecks, network capacity, etc. Through analytics, administrators and companies can more effectively manage existing computer system infrastructure while also more accurately planning for future needs.
In a data storage platform with multiple storage devices, performing analytics on the entire data store requires gathering information, such as metadata, from each node or storage device. For example, one approach to gathering stored data metadata is by scanning the storage platform from outside the storage platform using standard client access network protocols such as Network File System (“NFS”) and Common Internet File System (“CIFS”) protocols in a NAS file storage context and SCSI in a block storage context. However, using network protocols such as NFS and CIFS to request metadata from a storage platform by an external client can limit the metadata that can be gathered to metadata that can be expressed in the standard protocols. In addition, in some instances, the use of the network protocols in gathering metadata can negatively impact the performance of the data servers and/or storage platforms on which the metadata is being gathered. It can also increase network traffic, for example, by tree walking all network attached storage devices.